1. Field of the Invention
The present invention relates to a method of manufacturing photocathodes for use in photosensitive vacuum tubes, and more particularly, to a method for optimizing a photocathode's photo-response by means of a baking process.
2. Description of the Prior Art
It has previously been discovered that a layer of gallium arsenide (GaAs) deposited upon a photon transparent substrate and surface treated by a layer of cesium oxide forms an effective photocathode. Photocathodes of this type are utilized in image intensifiers, such as, those which are employed in night vision devices. In such intensifiers, the photocathode is typically deposited upon the inner surface of an input window which is vacuum sealed within a tube envelope. Electrons generated in the photocathode in response to an input of photons are accelerated and/or amplified by subsequent tube components to generate an intensified output image. For a photocathode to generate a photocurrent, i.e., a flow of electrons, in response to photon input, three things must take place. First, photons photon input, three things must take place. First, photons must be captured in the bulk material, e.g., GaAs, causing a release of electrons within the bulk material. Second, the released electrons must reach the surface of the bulk material before being reabsorbed. Lastly, the electrons must escape the attractive forces of the surface of the bulk material and proceed to the next component of the tube. It has been found that a cesium-oxide treatment of the surface of a GaAs layer reduces the attractive force on the photoelectrons in question so that a large portion of them escape, resulting in a large photo-response (PR). The amount of energy needed by an electron to leave the surface of a photocathode is called the work function. GaAs photocathodes that are properly activated with cesium-oxide actually have a negative work function and repel electrons from the surface into the tube. Imperfections in the cesium oxide surface treatment, such as molecules with the wrong stoichiometry, however, are attractive sites to the electrons and the regions immediately surrounding them have no PR, thereby reducing the overall PR of the device.
The cesium-oxide layer is typically deposited on the cathode surface by a vacuum deposition process. It has been found that the best photo-response results are achieved by monitoring the PR during the deposition process. In this manner, the process is terminated when the best compromise is achieved between maximizing the PR and incurring an excessive detrimental dark current. The deposition process is terminated when the PR is at its maximum level, however, the oxygen and cesium in the environment cannot be eliminated the instant maximum PR is sensed. Cesium, in particular, remains in the deposition environment longer and in greater quantity than oxygen. This causes excess cesium to be deposited on the surface of the photocathode, upsets the stoichiometry and degrades the PR. By the time the photocathode is sealed into the tube, the PR is less than 1/10 its previously observed maximum value. There are known processes, however, to restore the lost PR by baking the tube in an oven. It is also commonly accepted that baking an over-cesiated photocathode results in a photocathode having the most stable PR. Therefore, rather than avoiding over-cesiation, the standard practice in the industry is to over-cesiate the photocathode in the vacuum system and then subject the sealed tube to a baking process to restore the PR. It is the surface layer of cesium-oxide that is improved by the baking process rather than the GaAs layer.
The details of the physics of what happens during the PR bake process is not well understood, but it has been observed that the PR increases during the baking process, reaches a maximum, and then degrades if the process is continued. The PR lost by over-baking a tube is not recoverable. Tubes that are baked up to, but do not pass, their maximum PR also have maximum stability, i.e., the photo-response tends not to change during use. It is, therefore, important to sense when the PR is maximized and to stop baking at that time. Unfortunately, because the contribution to photo-response of the bulk material, e.g., GaAs, is adversely (but reversibly) affected by heating, the peak photo-response level of the surface layer can not be sensed merely by monitoring the photo-response of the tube during heating. The primary effect of heating on the bulk material is to make the electrons which are generated by the photon input reabsorb more readily before they can reach the surface to leave the photocathode. While the component of PR which is due to the cesium-oxide surface is maximized through baking, the bulk material component is diminished during the baking process. The PR measured during baking includes both the surface component and the diminished bulk material component. The adverse effects of the heat of baking on the PR of the bulk material layer reverse upon cooling, however, and the baking process does not permanently diminish the PR component attributable to the bulk material layer.
The rate at which the surface component of the PR increases as it is baked is controlled by the baking temperature and the proximity to maximized PR. The higher the baking temperature, the faster the increase in PR. The closer to the maximum achievable PR, the slower the increase. The rate of increase in surface component PR also depends upon the individual cathode.
There are presently two standard procedures for baking photocathodes to maximize PR. The first procedure is to bake the tube for a set period of time, e.g., one hour; let it cool to room temperature; and then measure the PR. This cycle is repeated over and over and the increase in PR is measured during each cycle. The bake temperature of each cycle is selected to keep the rate of increase in the desired range. When the rate of increase approaches zero, the baking process is terminated, this being an indication that maximum PR has been realized. There are certain disadvantages with this process, viz., each warming/cooling cycle consumes a great deal of time; there is a lack of control of the cool temperature which introduces error; and the warming and cooling times impose a practical limit both on the shortness of the bake time of each cycle and on the accuracy of finish time. The long time period between measurements also limits the number and resolution of the temperature increments. If the increments are infrequent, they must be large in order to arrive at the desired temperature.
Another common alternative is to bake the tube at a constant temperature. In this way, the PR can be measured at the bake temperature because the component of the PR which is affected by the temperature is held constant and only the component due to the surface layer is varying. This permits frequent measurements so the process can be stopped at the moment of peak PR. The disadvantage with this method is that there is no control of the rate of this process, i.e., by increasing the baking temperature. Also, because some tubes require higher than normal temperatures to bake properly, they cannot be maximized by this method.
It is therefore an object of the present invention to provide a method for baking a photocathode interactively varying the temperature of baking in response to measured PR.
It is a further object to provide a method for measuring PR while baking a photocathode that isolates the component of PR attributable to the surface layer from that of the bulk material, or, in other words, project the room temperature PR given a PR measurement while the cathode is being baked.